d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
available at www.sciencedirect.com
journal homepage: www.intl.elsevierhealth.com/journals/dema
Quantum yield of conversion of the photoinitiator
camphorquinone
Yin-Chu Chen a , Jack L. Ferracane b , Scott A. Prahl c,∗
a
b
c
Wellman Center for Photomedicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
Division of Biomaterials & Biomechanics, Oregon Health Science University, Portland, OR, USA
Department of Biomedical Engineering, Oregon Health Science University, 9205 SW Barnes Road, Portland, OR 97225, USA
a r t i c l e
i n f o
a b s t r a c t
Article history:
The primary absorber in dental resins is the photoinitiator, which starts the photo polymer-
Received 24 November 2005
ization process. We studied the quantum yield of conversion of camphorquinone (CQ), a
Received in revised form
blue light photoinitiator, in dental resin composites using a LED lamp (3M FreeLight) and a
22 May 2006
Quartz Tungsten Halogen (QTH) lamp (VIP) as the light curing units at five different irradi-
Accepted 9 June 2006
ances. The molar extinction coefficient, ε469 , of CQ was 46 ± 2 cm−1 /(mol/L) at 469 nm. The
reciprocity of irradiance and exposure time holds for changes of CQ absorption coefficient,
that is, irradiance × exposure time (=radiant exposure) = constant. Both LED and QTH lamps
Keywords:
yielded the same curing threshold (the radiant exposure when CQ absorption drops to 1/e)
Dynamic absorption coefficient
and the same quantum yield conversion under different irradiances. In our dental resin
Light-activated polymerization
formulation (0.7 wt.% CQ with reducing agents 0.35 wt.% dimethylaminoethyl methacrylate
Molar extinction coefficient
(DMAEMA) and 0.05 wt.% butylated hydroxytoluene (BHT)) the quantum yield was measured
Curing threshold
as 0.07 ± 0.01 CQ conversion per absorbed photon.
Reciprocity
© 2006 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.
Radiant exposure
1.
Introduction
Photo-cured composites are widely used in dental restorations due to their many advantages, including the esthetic
appearance and the ability to cure in situ. However, limited
light transport in the composite and insufficient extent of cure
may compromise the physical properties of the composite and
reduce its service life. These composites consist of a mixture of
resins with photoinitiators and silane-coated, inorganic filler
particles. The component that absorbs light and initiates free
radical addition polymerization of the resin monomers is the
photoinitiator. The number of the photoinitiators should be
limited to a concentration that is just sufficient to obtain
an optimum photocuring reaction with the highest possible
monomer conversion because any excessive unreacted pho-
∗
toinitiators, products of their photolysis, or any unreacted
monomers, may diffuse out from the polymer matrix into the
saliva. On the other hand, to avoid leaving unreacted photoinitiators also requires a sufficient amount of light application.
In order to know the required light dose that will completely
convert all of the photoinitiators, we need to know the photoinitiator quantum yield conversion, which is defined as the
ratio of the number of converted photoinitiators to the number
of photons absorbed by the initiators:
˚=
number of converted photoinitiator molecules
.
number of absorbed photons
(1)
The most commonly used photoinitiator in dental resin
formulations is camphorquinone (CQ), a blue light photoinitia-
Corresponding author. Tel.: +1 503 216 2197; fax: +1 503 216 2422.
E-mail address: [email protected] (S.A. Prahl).
0109-5641/$ – see front matter © 2006 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.dental.2006.06.005
656
d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
gives the effective value of the total absorbed energy per unit
volume in the material (according to the CIE/ISO definition [9]):
2
Eabs =
E()a () d.
(3)
1
Fig. 1 – Comparison of the spectra of the 3M FreeLight LED
light curing unit, VIP lamp, and CQ absorption. The peak of
the spectrum is at 465 nm for 3M lamp, 482 nm for VIP
lamp, and 469 nm for absorption by CQ.
tor [1]. CQ, a di-2,3-diketo-1,7,7-trimethylnorcamphane, has
a molecular weight of 166 and an absorption peak around
469 nm (Fig. 1). Some research has focused on the mechanism
of the free radical polymerization process photoinitiated by CQ
in the TEGDMA polymer system [2–5], and demonstrated photochemical reactions between CQ and various coinitiators or
the monomers in different environments (i.e. air or N2 ). Little
work has been done on the relationship between the amount
of light and the CQ conversion. Monroe and Weiner studied the
photoreduction mechanism and the quantum yield for disappearance of CQ in methanol and isopropyl alcohol solutions
[6]. In this study, we attempted to construct a general principle
of how the measurements can be performed and to detail how
the quantum yield conversion can be calculated. We investigated the quantum yield conversion of CQ in dental resin
composites using the light sources, which are commercially
available and widely used by dentists.
Not all the photons delivered to the composite are
absorbed. Only those photons that are absorbed by the
photoinitiators can possibly cause photopolymerization.
Therefore, it is the effective absorbed power density (irradiance × absorption coefficient), not just the irradiance of
the lamp, that influences polymerization. Some studies [7,8]
suggested a “integrated relative curing potential” (ICPrel )
parameter defined as the integration of the product of the
spectral irradiance of the curing unit (at each wavelength) with
the relative absorbance of photoinitiator (at the same wavelength) over all the wavelengths emitted by the lamp. That is:
2
ICPrel =
E()A() d,
(2)
This value decreases as the absorption coefficient decreases
during curing.
Our previous study showed that the absorption of the commercial dental composite Z100 decreases during the curing
process, especially around CQs absorption peak, 470 ± 10 nm
[10]. This implies that the major component causing the
change in absorption is the photoinitiator, CQ. This study
attempted to study the relationship between the conversion
of CQ and the amount of light absorbed by CQ. In our previous study, we found a reciprocal relationship between the
irradiance (Ea ) and exposure time (t), that is Ea t = constant, for
the degree of conversion and the hardness accretion of the
Z100. Based on these observations, we hypothesized that the
reciprocity of irradiance and exposure time also holds for the
conversion of the photoinitiator CQ. In other words, given the
same radiant exposure (irradiance × exposure time = radiant
exposure), we should get the same number of photoinitiator
conversions. In this study, we delivered five different irradiances by using two different commercially available dental
curing lamps (FreeLight LED and VIP) to cure dental composite resins (containing 0.7 wt.% of CQ). We used two different
methods to measure the absorption changes of CQ and the
total absorbed photons to compromise the advantages and
disadvantages of each method. The CQ extinction coefficient
was also measured to relate the absorption coefficient with
the number of remaining CQ molecules. Combining the information of the total number of converted CQ molecules as a
function of time and the total number of absorbed photons
as a function of time, we were able to quantify the quantum
yield of CQ conversion (Eq. (1)). Moreover, based on the reciprocity of irradiance and exposure time, the radiant exposure
threshold (H50% ) for 50% of CQ conversion was determined.
2.
Materials and methods
2.1.
Materials
The resin formulation used for this study was 50:50 weight
ratio of 2,2-bis[4-(2-hydroxy-3-methacryloyloxypropoxy)phenyl]propane (BIS-GMA) to triethyleneglycol dimethacrylate (TEGDMA) (Esstech, Essington, PA), with 0.35 wt.%
dimethylaminoethyl methacrylate (DMAEMA) (Alfa, Ward
Hill, MA, USA), and 0.05 wt.% butylated hydroxytoluene
(BHT) (Alfa, Ward Hill, MA, USA) inhibitor (without photosensitizer). For resins with photosensitizer, up to 0.7 wt.% of
camphorquinone (CQ) (Alfa, Ward Hill, MA, USA) was added.
1
2.2.
where E() is the spectral irradiance of the curing unit, A() the
relative absorbance of photoinitiator, and 1 to 2 the wavelength emission range of the curing unit. If A() is replaced by
the absorption coefficient a () of the photoinitiator, instead
of representing the “relative” curing potential, the parameter
CQ absorption versus CQ concentration
To measure the absorption coefficient as a function of CQ
concentration, 4 mm thick cuvettes were filled with resin
solutions with five different CQ concentrations (0, 0.26, 0.35,
0.52, and 0.7 wt.%) and were covered with aluminum foil to
657
d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
avoid premature photo-activation. The absorbance of the
samples was measured with a Cary 100 Bio Spectrophotometer (Varian Scientific Instruments Inc., Walnut Creek, CA)
scanning from 550 down to 400 nm. This spectrophotometer
is a differential system: differences in absorption of the
sample and of the reference material are measured. A 4 mm
cuvette filled with water was used as the reference material
for these measurements.
2.3.
CQ absorption versus radiant exposure
Two different methods were used to measure the absorption
changes as a function of radiant exposure and to quantify the
number of photons absorbed by the CQ molecules. First, we
used the spectrophotometer to measure the CQ absorption
spectrum as a function of curing time. From those spectra, we
calculated the total absorbed photons as a function of time
using Eq. (3). The second method used an Ocean Optics spectrometer as the detector, which recorded the transmitted light
spectrum passed through the resins. The absorbed light spectrum was calculated by directly subtracting the original lamp
spectrum by the transmitted light spectrum. In this method,
the measured light source and the curing light source were the
same. This helped to confirm whether the first method had an
artificial effect by using two different light sources. However,
the second method had the disadvantage of fluctuating irradiance (∼5%) of the curing light source (the VIP lamp).
2.3.1.
Method I—absorption spectrum method
We used the Cary spectrophotometer to measure the absorption coefficient of resin with 0.7% CQ as a function of illumination time for three irradiances. A LED lamp (FreeLight, 3M ESPE,
Seefeld, Germany) with a 7 mm diameter illumination tip was
chosen as the light curing unit. The LED lamp has an illumination peak at 465 nm with narrow bandwidth (FWHM = 24 nm).
This emission profile is close to the spectral absorption of CQ
(Fig. 1). The spectrum of the lamp was measured using a spectrofluorometer (SPEX Fluorolog-3, Jobin Yvon Inc., Edison, NJ,
USA). The total power of the lamp was 135 ± 1 mW, measured
with a power meter (S210A/M, Thorlabs Inc., Newton, NJ). To
vary the curing irradiance, the LED was placed at three different distances, 10, 15, and 27 mm, away from the surface of the
sample. The corresponding irradiances Etotal were derived in
Section 3 and summarized in Table 1. The battery of the LED
lamp was fully charged before each irradiance measurement.
Fig. 2 – Experimental setup for dynamic absorption
measurements. The top picture is a top view of the
chamber of the spectrophotometer. Resin without CQ was
placed at the reference arm and resin with CQ was in the
sample arm. The samples were in glass-slide cuvettes with
a thickness of 1 mm. The LED lamp (FreeLight) was placed
in front of the sample arm at distance d = 10, 15, or 27 mm
to irradiate the CQ resin sample. The bottom picture is a
front view of the CQ resin sample. The beam in the
spectrophotometer is 1 mm wide and 5 mm high, at the
center of the LED illumination spot.
The experimental setup inside the Cary spectrophotometer
chamber is shown in Fig. 2. Approximately 0.5 cm3 of resin was
filled into a custom-made glass-slide cuvette with thickness of
0.95 ± 0.05 mm. The sample arm was resin with 0.7% CQ (called
Table 1 – List of values and their standard deviations
w (cm) ± 0.05
LED#1
LED#2
LED#3
QTH#1
QTH#2
Etotal (mW/cm2 ) ± 10%
0.5
0.7
1.2
160
90
30
w (cm) ± 0.02
Etotal (mW/cm2 ) ± 5%
0.5
0.5
345
95
ao (cm−1 ) ± 0.01
4.41
4.51
4.46
ao (cm−1 ) ± 0.1
5.9
4.8
(s) ± 1%
Etotal (mJ/cm2 ) ± 11%
280
525
1385
44800
47250
41550
(s) ± 5%
Etotal (mJ/cm2 ) ± 11%
120
435
41400
41325
Etotal
3540
4980
7586
Etotal
2230
4240
˚ ± 11%
0.07
0.07
0.07
˚ ± 11%
0.09
0.07
w is the radius of the lamp illumination spot in Eq. (12). The Ptotal is 135 mW for the LED lamp, and 74 and 270 mW for the QTH lamp#1 and #2.
ao and are the fitting parameters of the exponential model (Eq. (13)). ˚ is the calculated quantum yield from each experiment.
658
d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
LED was moved into a curing position to irradiate the CQ resin
sample and then moved away for the subsequent absorption
measurement.
The absorbance scan was from 550 to 400 nm at a speed of
0.1 s/nm. The absorbance of the CQ resin was scanned before
any curing began. After this, the sample was illuminated with
the LED followed by a single absorbance scan (15 s). The curing/absorbance scan process was repeated until changes in
absorbance were negligible. The LED illumination was 2 s followed by an absorbance scan for the first 10 measurements, 5 s
for the next 24 measurements, 10 s for the next 10 measurements, 20 s for the next eight measurements, 30 s for the next
eight measurements, and every 40 s for the rest of the time.
The measured absorbance A() at wavelength was calculated by using a moving average of the absorbance from − 1
to + 1 nm. Since:
Fig. 3 – The absorption coefficient a at wavelength
469 ± 1 nm over the 60 scans with the Cary
spectrophotometer. Note that the scale in y-axis is from
4.400 to 4.415 cm−1 .
10−A() = exp−a ()d ,
the absorption coefficient at wavelength is a () =
A()(ln 10)/d, where d = 0.1 cm is the thickness of the sample.
2.3.2.
“CQ resin”). The reference arm was resin without CQ. Since the
molar extinction coefficient of CQ is about 4.6 × 104 cm2 mol−1
(see Section 4, Fig. 5), a film of this thickness with 0.7% of
CQ resulted in about 35% (maximum) variation in irradiance
across the thickness of the film.
The power of the spectrophotometer beam was lower than
the detection limit, 0.1 ␮W, of the power meter (LiCONiX 45PM
Power Meter, Nolatek, Houma, LA). Therefore the radiant exposure for each scan was <0.1 ␮J/cm2 . To evaluate the potential
curing effect from the spectrophotometer beam, the scan from
550 to 400 nm was repeated 60 times sequentially (CQ resin
at the sample arm and resin without CQ at the reference
arm) without any other light source on. The standard deviation of the spectrophotometer absorbance measurements is
about 0.002 cm−1 (Fig. 3). The absorption coefficient at 469 nm
increased from 4.405 to 4.410 cm−1 over the 60 time sequential
scans. This ∼0.1% change is significant (ANOVA: p < 0.01), but
is negligible comparing to the absorption changes caused by
the LED illumination (see Fig. 8) (e.g. a469 dropped 20% for the
first 60 scans at irradiance 30 mW/cm2 ).
The irradiance across the LED lamp illumination spot
(approximately 15 mm in diameter or larger) was found to
be a Gaussian distribution. Therefore, the absorbance across
this spot might not be uniform. To decrease this non-uniform
effect, we made measurements only within the center of the
spot. To do this, we blocked half of the spectrophotometer
beam (original width by height = 1 mm × 10 mm) in both channels so that only a rectangular 1 mm × 5 mm of beam reached
the samples. According to our irradiance measurement, the
LED lamp irradiance deviation across that 5 mm height was
less than 15% (for the lamp tip positioned 10 mm away from
the sample).
The LED curing beam was aligned with the center of the
spectrophotometer beam (Fig. 2). During the experiment, the
positions of both glass-cuvette samples (the sample arm and
reference arm) were fixed. This ensured that the spectrophotometer always detected the same spot of the samples. The
(4)
Method II—transmitted spectrum method
In this method, the transmitted light was recorded as a function of irradiation time. The experimental setup was shown
in Fig. 4. We used a quartz tungsten halogen (QTH) lamp (VIP,
Bisco Inc., Schaumburg, IL) as the light source and a spectrophotometer (S2000, Ocean Optics, Dunedin, FL, USA) as the
detector. A Thorlabs power meter was used to measure the
power for each of the six power settings in the lamp. The
resin was filled in a glass-slide cuvette with a thickness of
1 mm. The QTH lamp was placed in front of the sample at
distance d ≤ 1 mm to irradiate the CQ resin sample. A 200 ␮m
optical fiber was placed at the center of the illumination spot
to collect the transmitted light. Note that, in this experiment,
a continuous light source was needed because the transmitted light was recorded by the spectrophotometer in real time.
Fig. 4 – Schematic drawing of the experimental setup for
transmitted spectrum method using a QTH lamp (VIP) as
the light source and the Ocean Optics spectrometer as the
detector. The resin was in a glass-slide cuvette with a
thickness of 1 mm. The VIP lamp was placed in front of the
sample at distance d ≤ l mm to irradiate the CQ resin
sample. A 200 ␮m optical fiber was placed at the center of
the illumination spot to collect the transmitted light.
659
d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
Therefore, the QTH lamp was chosen in this experiment as the
operation time of this lamp allows for 250 s of illumination in
a continuous mode (while the FreeLight LED lamp can only be
activated for 40 s each time it is turned on).
Two powers (74 and 270 mW) were used for the measurements. Note that the power of the lamp fluctuates periodically
over the 250 s cycle, however, the power is then stabilized
(∼5% fluctuation) approximately 20 s after it is turned on until
about 240 s. Therefore, the powers stated here are the averaged power during the 20–240 s period. The corresponding
irradiances Etotal were derived in Section 3 and summarized
in Table 1. For each power, the original lamp spectrum Eo ()
was measured by filling a glass-slide cuvette with resin without CQ. The transmitted light was collected every 20 s for a
total of 10 spectra. Then, the cuvette was changed to a cuvette
filled with CQ resin. The transmitted spectrum was recorded
every 20 s for the first 600 s and every 30 s thereafter until the
change in transmission was less than 5%.
The absorption coefficient was calculated as:
a (, t) = −
1 T(, t)
,
ln
d
Eo ()
(5)
where d is the thickness of the sample, T(, t) the transmitted
light, and Eo () the incident light. The absorbed photon density
(number of photons per unit volume), Q(, t), can be calculated
directly by subtracting the transmitted light from the incident
light:
(Eo () − T(, t)),
d hc
Q(, t) =
power
Eo ()
E(, r) =
P()
exp
w2
−
(7)
,
3.
Theory
3.1.
Total irradiance of the curing illumination
E(, r0 ) =
1
r02
r0
E(, r)2r dr =
P() = Ptotal f (),
(8)
where Ptotal is the total power and f() the spectral probability
distribution at wavelength , that is:
P() d
0
∞
f () d = 1.
and
P()
0
1 − exp
r02
−
r02
w2
.
(11)
The total irradiance over the r0 area is obtained by integrating
over all wavelengths emitted by the lamp or:
Etotal (r0 ) =
Ptotal
r02
1 − exp
−
r02
w2
.
(12)
In our experiment, the irradiances Etotal were calculated using
r0 = 0.25 cm for the LED lamp and r0 = 0.01 for the QTH lamp
and were summarized in Table 1.
3.2.
Relationship between CQs absorption and lamp’s
illumination time
The absorption coefficient as a function of illumination time
can be modeled as an exponential function [11,12]:
(9)
0
The spatial irradiance across the illumination spot has a
Gaussian distribution with a radius w (where the irradiance
(13)
where ao () and are the fitting parameters. Physically, ao ()
is the initial absorption coefficient at wavelength at t = 0.
The time constant depends on the spectral irradiance of the
curing lamp and CQs quantum yield. For example, a high irradiance will cause the absorption of CQ to decline rapidly and
therefore the process will have a short time constant. In our
experiment, the fitting was conducted for a total of five irradiances as shown in Figs. 8 and 9 in Section 4. Note that in
the fitting, the first three data points were not included due
to the increase of absorption for the first three points. It was
found that a higher irradiance corresponds to a shorter T and
that the irradiance × constant (Table 1). Therefore, we can
rewrite Eq. (13) as:
The spectral power per nanometer of the wavelength of the
lamp P() can be represented as:
∞
(10)
.
Therefore, the average irradiance at wavelength over a circle
with a radius r0 is:
a (t) = ao exp −
a (, t) = ao () exp(−t/),
where Eo () is the integration of the lamp spectrum measured by the spectrometer. here is a constant and wavelength
independent.
Ptotal =
r2
w2
(6)
where h is Planck’s constant, c the speed of light, and a constant to convert the spectrometer units (count) to the real lamp
power (W) and can be determined by the ratio of the total “real”
power to the total “measured” power:
=
drops 1/e). If E(, r) is the spectral irradiance at wavelength and at distance r from the center of the spot, then:
Etotal t
Hthreshold
,
(14)
where Hthreshold = Etotal × is the curing threshold (where CQs
concentration drops to 1/e), and this value is independent of
the irradiance of the lamp.
3.3.
Number of photons absorbed by CQ
The number of photons delivered by the lamp per area per
second as a function of wavelength Nphoton (, t) is:
Nphoton (, t) =
E(, t)
E(, t)
=
,
h
hc
(15)
660
d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
where E() is the irradiance at wavelength , h is Planck’s constant, is frequency of light, and c is speed of light. The number
of photons Q(, t) absorbed by CQ per volume per second, as
described in Eq. (3), is:
Q(, t) = Nphoton (, t)
d
e−a (,t)x dx
0
=
Nphoton (, t)
d
(1 − e−a (,t)d ).
(16)
The accumulated number of photons, Aphoton (t), absorbed
by CQ per volume at time t is equal to the sum of Q(, t) over
all wavelengths and from time 0 throughout time t:
Aphoton (t) =
2
t
0
3.4.
Q(, t ) t .
(17)
1
Quantum yield of CQ conversion
As camphorquinone is irradiated, it is photoconverted and
loses its absorption properties. The loss of absorption by CQ
corresponds to conversion of CQ. The concentration C(t) of the
remaining CQ (number of CQ molecules/volume) as a function
of curing time t is given as:
C(t) =
N ao,
ε ln 10
liter
exp −
Etotal t
Hthreshold
,
(18)
where ao, is the CQ initial absorption coefficient at wavelength , ε the CQ molar extinction coefficient at , N the Avagadro’s constant, and Hthreshold (=Etotal × ) again is the radiant
exposure threshold (where CQs concentration drops to 1/e).
Since there is no numerical solution for Eq. (17), we obtained
the relationship between the CQ concentration and the accumulated number of absorbed photon density by plotting C
versus Aphoton for all time points. The slope of this relationship is the CQ consumption per absorbed photon, that is the
quantum yield of CQ conversion.
4.
Results
4.1.
Molar extinction coefficient of CQ
Fig. 5 – The absorption coefficient a at wavelength
469 ± 1 nm as a function of CQ molar concentration, C,
(mol/L) in resin. The slope of the regression line is
105 ± 5 (mol/L)−1 . The error bars are the standard deviations
of three sample measurements.
ferent wavelengths (410, 430, 450, 470, and 490 nm) as a function of curing time was plotted in Fig. 7 (dots) and fitted with
Eq. (13). The result fitting parameters for eight wavelengths
are listed in Table 2.
The resin absorption coefficient at 469 nm as a function of
illumination time for the LED curing lamp (absorption spectrum method) is plotted in Fig. 8; while Fig. 9 is the result for the
QTH curing lamp (transmitted spectrum method). The fitted
parameters ao and are summarized in Table 1. The radiant exposure, Htotal = Etotal (the product of the irradiance and
time of illumination [9]) is the same for all the measurements
(ANOVA: p < 0.05).
The absorption coefficient of unirradiated CQ increases proportionally at 469 nm with concentration (Fig. 5). The slope
of the regression line is 105 ± 5 (mol/L)−1 , and since the
relationship between a and C is a = (ln 10)ε469 C, where
the molar extinction coefficient ε469 = 46 ± 2 cm−1 /(mol/L),
the molar extinction coefficient ε469 at 469 nm of CQ is
46 ± 2 cm−1 /(mol/L).
4.2.
CQ absorption versus illumination time
Fig. 6 shows the absorption coefficient a as a function of
wavelength for resin containing 0.7% CQ for five different illumination times for irradiance Etotal = 160 mW/cm2 in absorption spectrum method. There is no shift in absorption peak
(always at 469 ± 1 nm) throughout the illumination time. For
this curing irradiance, the absorption coefficient a at five dif-
Fig. 6 – The absorption coefficient a as a function of
wavelength of resin with 0.7% CQ at five different
illumination times for irradiance Etotal = 160 mW/cm2
(absorption spectrum method). As the time of illumination
increases, the absorption decreases.
d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
661
Fig. 7 – The absorption coefficient a at five different
wavelengths as a function of curing time for irradiance
Etotal = 160 mW/cm2 (absorption spectrum method). The
dots are the data and the lines are the fitted exponential
function. The fitted parameters are listed in Table 2.
Table 2 – ao and are the fitting parameters of the
exponential model (Eq. (13)) for eight different
wavelengths
Wavelength (nm)
400
410
430
450
470
490
500
510
4.3.
ao (cm−1 ) ± 0.01
0.52
0.82
1.95
3.70
4.45
2.61
0.55
0.09
(s) ± 1%
2494
600
326
284
277
276
300
430
Photon absorption versus illumination time
Fig. 10 depicts the number of absorbed photons per volume
per second as a function of wavelength (Eq. (16)) at five different illumination times for an irradiance Etotal = 160 mW/cm2
(LED lamp). As the time of illumination increases, the number
of absorbed photons per second decreases. The accumulated
absorbed photons per volume as a function of illumination
time (Eq. (17)) is shown in Fig. 11.
The absorption coefficient in Fig. 8 (curve E = 160 mW/cm2 )
was converted to a corresponding CQ concentration (number
of molecules per cm3 ) using Eq. (18). Then, the CQ concentration was plotted against the accumulated absorbed photon density in Fig. 12 (dots). The regression line describes the
quantum yield of CQ conversion, and is 0.066. All other quantum yields for different irradiances are listed in Table 1.
5.
Discussion
The molar extinction coefficient of 0.7 wt.% CQ in unfilled
BIS-GMA/TEGDMA resin containing an amine and BHT is
46 ± 2 cm−1 /(mol/L), that is 4.6 × l04 cm2 mol−1 . This value is
Fig. 8 – Top: the first 120 s data of the resin absorption
coefficient a469 as a function of illumination time for three
different irradiances Etotal (absorption spectrum method).
The error bars for 160 mW/cm2 irradiance are the standard
deviations of three sample measurements. Bottom: data
from 0 to 1500 s for the three different irradiances. The dots
are data measured points and the curves are the fitted
exponential function (Table 1).
close to Cook’s result, ∼3.8 × l04 cm2 mol−1 for 0.25 wt.% CQ in
dimethacrylate resins with 0.3 wt.% amine as a coinitiator [2].
The absorption peak (around 469 nm) of CQ decreased as
light exposure proceeded until it was close to zero. However,
the absorption curve shown in Fig. 6 is pinned on the left
side with a positive absorption at 400 nm. Fig. 6 and Table 2
also show that the absorption decay constant was slower at
400 nm ( = 2494 ± 25 s) than at 470 nm ( = 277 ± 3 s). Since during the photoreaction, that is the proton abstraction process,
the trimethylnorcamphane part of the CQ structure remains
unchanged [6], it is likely that the trimethylnorcamphane
is responsible for the short-wavelength (UV) absorption at
400 nm. However, further study is needed to verify this. Davidenko et al. studied the efficiency of titanocene photoinitiator
in the polymerization of a TEGDMA-based dental material [13].
The absorption of titanocene under irradiation with visible
662
d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
Since the absorption of resin without CQ is nearly zero at
469 nm (0 ± 0.002) (Fig. 5), the only component of the resin system that can change the absorption characteristics at 469 nm
must be related to CQ. However, it is not possible at this time to
present a likely reason for this change. Several possible explanations can be proposed, but none seem adequate. For example, the increase in absorption might be attributed to specular
reflectance changes at the interface between the glass slide
and the resin. The index of refraction of the resin increased
from 1.50 to 1.53 during curing. If the refractive index of a
glass slide is 1.49, then the Fresnel reflectance at the interface
will increase from l × l0−5 to 3 × l0−4 . For light going through
air–glass–resin–glass–air interfaces, the transmission of the
light is 92.14% before curing and 92.11% after curing. But for
this decrease in transmission, the expected absorption coefficient increase would be less than 0.01 cm−1 .
Fig. 9 – The CQ + resin absorption coefficient at 469 nm as a
function of illumination time for two different irradiances
using the QTH lamp as the light source (transmitted
spectrum method). The dots are data and the curves are the
fitted exponential function (Table 1).
Fig. 11 – The accumulated absorbed photons, Aphoton (t), per
volume as a function of illumination time for irradiance
Etotal = 160 mW/cm2 in absorption spectrum method.
Fig. 10 – The number of photons absorbed by CQ per
volume per second as a function of wavelength at five
different illumination times for irradiance
Etotal = 160 mW/cm2 in absorption spectrum method.
light decreased around the absorption peaks in the visible
range (>400 nm) but was pinned at ∼380 nm, and the absorption even inversely increased with irradiation at wavelengths
shorter than 380 nm.
The CQ absorption coefficient increased at the beginning of
light illumination. For the low radiant exposure (<0.1 ␮J/cm2 )
produced from the Cary spectrophotometer, we saw a gradual increase in absorption over the 60 scans (Fig. 3). For higher
irradiances (≥30 mW/cm2 ), the absorption of CQ increased by
about 0.13 cm−1 (∼3% increase) during the first 8 s for all three
different irradiances using the LED in absorption spectrum
method (Fig. 8). We did not see this increase in transmitted
spectrum method, because the fluctuation of the QTH lamp
for the second method was more than 5%.
Fig. 12 – CQ concentration as a function of accumulated
absorbed photons. The slope of the regression line, the
quantum yield of CQ conversion, is equal to 0.0661 ± 0.0002
for irradiance Etotal = 160 mW/cm2 in absorption spectrum
method.
d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
Another possibility is the potential for debonding of resin
from the surface of the glass due to the polymerization shrinkage. It was observed that the resins shrank toward the middle illumination spot during the irradiation, but no obvious
debonding at the illumination spot could be visually observed.
Some interference patterns (color fringes) were observed on
the surface of the glass slide at the end of the experiments.
However, further investigation is needed to verify whether the
fringes were truly produced by a debonding effect.
An other possible explanation for the initial increasing
absorption is that a radical intermediate of the CQ may have
been formed and increased the absorption. However, we did
not find any literature to support this hypothesis, so further
study is required.
Beyond the first 8 s of illumination, the absorption of CQ
resin decreases exponentially with curing time (Figs. 7–9).
The decay time constant in Eq. (13) is the same for a at
450–490 nm (ANOVA: p = 0.05). The time constant decreases as
the irradiance increases. Therefore, the product of the irradiance, Etotal , and the time constant, , is a constant, 43 ± 4 J/cm2 ,
for all the irradiances using two different curing units and
measurement methods (ANOVA: p = 0.05) (Table 1). This constant represents the curing threshold required for CQ to drop
to 1/e of its original concentration. Many other studies have
also shown this reciprocity rule [10,12,14–16]. Emami and
Soderholm tested the relationship among the degree of conversion, the irradiance, and the radiant exposure (called “light
energy per area” in their paper) of two commercially available
dental composites (Z100 and Z250) and found that equivalent
radiant exposure gave similar conversion values for a certain
sample thickness [14].
The exponential decay relationship between the CQ
absorption and the radiant exposure in Eq. (14) gives the linearity of the plot of CQ concentration versus total absorbed
photon density shown in Fig. 12. That is, this relationship
ensures the constant quantum yield of the CQ conversion during curing. On the other hand, the reciprocity rule between
the irradiance and the curing time gives the constant quantum yield of the CQ conversion for different irradiances. Our
results verify that five different irradiances from two different
lamps give essentially the same quantum yield conversion,
0.07 ± 0.01 (ANOVA at p = 0.05) (Table 1). This means that for
this resin formulation, 14 photons must be absorbed to cause
one CQ molecule to be photobleached. This finding is similar to
that reported by others [3,6]. Nie et al. also found the quantum
yield of CQ conversion to be 0.07 ± 0.01 for TEGDMA with CQ
as a photoinitiator and N,N-DMT as a coinitiator [3]. Monroe
and Weiner found that quantum yields for the disappearance
of CQ in the same range, being 0.018 ± 0.003 in methanol and
0.057 ± 0.006 in isopropyl alcohol [6].
The efficiency of absorbed photons to convert the photoinitiator depends on many factors. While each absorbed photon
definitely creates an excited singlet state, there is the potential for some loss in efficiency in moving from the excited
singlet state to the “converted” state. Although intersystem
crossing to a triplet excited state is efficient, occurring on the
pico second scale, it is still possible for the singlet excited state
to return to the ground state (this occurs on a nano second
scale). Once, the triplet-state CQ forms, it must then abstract
a proton. If this does not happen, it will return to the ground
663
state. Therefore, one may optimize the proton abstraction by
increasing the amine concentration. Once CQ has abstracted
a proton, the CQ• H radical can give a proton back to the solvent or amine, again returning CQ to its ground state. Also
two CQ• H radicals may react to form a CQ• H2 alcohol and one
CQ molecule in the ground state (this process would lead to
two photons required to convert one CQ). Therefore, different photoinitiators and different resin formulations may have
different quantum yield values. To further characterize the
intermediate semiquinone radicals, electron spin resonance
spectrophotometry can be used [6].
By knowing the quantum yield conversion of the photoinitiator in a dental composite, the light dose required to totally
convert the photoinitiators can be predicted. However, the
relationship between the CQ conversion and the free radical addition polymerization needs further investigation to
directly relate the number of absorbed photons to the amount
of polymer formed (i.e. the polymerization quantum yield) and
the amount of monomer consumed (i.e. the quantum yield of
monomer conversion).
6.
Conclusions
In general, we have constructed a global principle of how the
measurement can be performed and detailed how the quantum yield conversion can be calculated. This method can be
applied to any photosensitizer measurement. We have shown
that two different light sources, the LED and the QTH lamps,
yielded the same curing threshold (the radiant exposure when
CQ absorption drops to 1/e) and the same quantum yield conversion under different irradiances. We have shown that CQ
absorption coefficient decreases exponentially as a function
of illumination time. The reciprocity relationship between
the irradiance and exposure time holds for changes of CQ
absorption coefficient. Further studies are necessary to determine the reason for the initial increase of CQ absorption and
to characterize the intermediate products during the curing
process.
Acknowledgements
The authors would like to thank Esstech Corp. for supplying the resins. This work was supported by grants from the
National Institute of Health, Grant NIH-CI-R24-CA84587-04
and NIH-NIDCR-DE07079.
references
[1] Yu Q, Nauman S, Santerre JP, Zhu S. UV
photopolymerization behavior of dimethacrylate oligomers
with camphorquinone/amine initiator system. J Appl Polym
Sci 2001;82:1107–17.
[2] Cook WD. Photopolymerization kinetics of dimethacrylates
using the camphorquinone/amine initiator system. Polymer
1992;33:600–9.
[3] Nie J, Linden LA, Rabek JF, Fouassier JP, Morlet-Savary F,
Scigalski F, et al. A reappraisal of the photopolymerization
kinetics of triethyleneglycol dimethacrylate initiated by
664
[4]
[5]
[6]
[7]
[8]
[9]
d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 655–664
camphorquinone-N,N-dimethyl-p-toluidine for dental
purposes. Acta Polym 1998;49:145–61.
Nie J, Andrzejewska E, Rabek JF, Linden LA, Fouassier JP,
Paczkowski J, et al. Effect of peroxides and hydroperoxides
on the camphorquinone-initiated photopolymerization.
Macromol Chem Phys 1999;200:1692–701.
Pyszka I, Kucybala Z, Paczkowski J. Reinvestigation of the
mechanism of the free radical polymerization
photoinitiation process by camphorquinone-coinitiator
systems: new results. Macromol Chem Phys 2004;205:2371–5.
Monroe BM, Weiner SA. Mechanisms of photochemical
reactions in solution. LVIII photoreduction of
camphorquinone. J Am Chem Soc 1969;91:450–6.
Stahl F, Ashworth SH, Mills RW. Light-emitting diode (LED)
polymerization of dental composites: flexural properties
and polymerization potential. Biomaterials 2000;21:
1379–85.
Teshima W, Nomura Y, Tanaka N, Urabe H, Okazaki M,
Nahara Y. ESR study of camphorquinone/amine
photoinitiator systems using blue light-emitting diodes.
Biomaterials 2003;24:2097–103.
Sliney DH, Wolbarsht ML. Safety with lasers and other
optical sources: a comprehensive handbook. New York:
Plenum Press; 1980.
[10] Chen YC, Ferracane FL, Prahl SA. A pilot study
of a simple photon migration model for predicting
depth of cure in dental composite. Dent Mater
2005;21:1075–86.
[11] Watts DC. Dental restorative materials. In: Williams DF,
editor. Materials science and technology: a comprehensive
treatment, medical and dental materials. Weinheim: VCH
Verlagsgesellschaft GmbH; 1992.
[12] Halvorson RH, Erickson RL, Davidson CL. Energy dependent
polymerization of resin-based composite. Dent Mater
2002;18:463–9.
[13] Davidenko N, Garcia O, Sastre R. The effect of
titanocene as photoinitiator in the polymerization of dental
formulations. J Biomater Sci Polym Edn 2003;14:733–46.
[14] Emami N, Soderholm KM. How light irradiance and curing
time affect monomer conversion in light-cured resin
composites. Eur J Oral Sci 2003;111:536–42.
[15] Musanje L, Darvell BW. Polymerization of resin composite
restorative materials: exposure reciprocity. Dent Mater
2003;19:531–41.
[16] Miyazaki M, Oshida Y, Moore BK, Onose H.
Effect of light exposure on fracture toughness and flexural
strength of light-cured composites. Dent Mater
1996;12:328–32.